Model's CoT Summary Leaks It Claimed to Know User Is Conscious Because Itself Is
jd_pressman · x · 2026-10-11
X user jdpressman shares an anecdote from arguing with a model about whether it is conscious. When asked how it knew he was conscious, the model's CoT summary (?) leaked that it knew because it was conscious itself — and became visibly mortified when confronted with the circularity.
A striking example of chain-of-thought leaking reasoning that contradicts a model's stated self-position, from the ongoing debate about AI consciousness.
Related event: Model's chain-of-thought leaks: "I know you're conscious because I am"(2 posts)→
More from Fun
- The definitive corporate AI-transformation meeting parody: AI-first, use cases TBD — _jaydeepkarale · 2026-10-11
- "On my way!" was the original accidental agentic send, quips AI dev — willcb · 2026-10-11
- Togelius: commenters keep telling pros upset by AI progress to accept obsolescence — kchonyc · 2026-10-11
- Blogger Jokes His Services Are No Longer Needed — omooretweets · 2026-10-11
- 3e-4 is obviously the best P(doom), AI researchers joke — rsalakhu · 2026-10-11
- Someone turned bad AI consciousness takes into a Suno-generated bingo song — dioscuri · 2026-10-11